A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Transportation Engineering, School of Engineering and Technology

Thesis (M. Eng.) - Asian Institute of Technology, 2010

A routing solution for delivery systems with a heterogeneous vehicle fleet was developed to address constraints and improve efficiency. The solution was designed as a prototype to assist in strategic and operational decisions in organizations. In the model, different types of vehicles were located at a depot. A set of geographically spread customers were served by a heterogeneous fleet of vehicles. The demand of the customers was known from orders. The objective was to determine the shortest travel distance for each route to reduce total cost while still satisfying customers' requirements. Each customer was visited by only one vehicle. A heuristic algorithm based on genetic algorithms was chosen to address the calculation time limit which in reality cannot be met by traditional optimization methods. The algorithms were designed and tested for several parameters. A genetic algorithm involves three main procedures. Initially, permutation encoding is used as format to construct a set of proto-chromosomes (parent). Second, a crossover procedure was applied to select two proto-chromosomes as parent strings for crossover and to define crossover positions. The final procedure was a mutation operation based on mutation probability. The results showed an optimized route algorithm that produced the best quality within a limited time compared to a cluster based heuristics method and the real route table from the company the data were derived from. The genetic algorithm's performance was evaluated and appropriate values of parameters were defined.